The proposed research investigates procedures for rapidly teaching new relational discriminations, including basic communication skills, to mentally retarded children and young adults. The goal is to develop a systematic, efficient basis for generating and expanding behavioral repertoires, and to provide methods for increasing effective participation by these individuals in skill development programs. Rule-governed behavior and receptive communication within the matching-to-sample paradigm receive emphasis. The research concentrates on determining whether mentally retarded individuals can perform according to generalized principles that will help them learn relations between spoken words or visual symbols and objects, pictures, or other environmental events. In particular, the question is asked whether retarded subjects will respond in ways that resemble elementary logical operations, or can learn to do so, based on their cumulative experience. The focus is analyzing procedures that may generate stimulus control described by logical rules and that may allow retarded subjects (1) to respond in orderly and predictable ways with novel stimuli or stimulus configurations, and (2) to learn constructive and appropriate new behavior as a result of this experience. For broad relevance, it is proposed to study a range of mental retardation levels, including individuals that are difficult to study and teach using traditional methods. Results of these studies are likely to have general theoretical and educational significance. By evaluating when and what types of "rules" are learned, and by exploring the limits of their application, it may be possible to contribute to the design of more effective teaching strategies and programs.